Mean opinion score

MOS is a speech quality metric that has been in use in telecommunication for years as a tool to analyze human users’ perceptions of call quality. The test is commonly used in VoIP networks to assure high-quality voice transmission, detect quality problems, and offer a metric for evaluating voice deterioration and performance. With the growth in popularity of phone services, MOS rating is becoming increasingly important for assuring customer happiness and network expansion.

Measure and metrics for MOS

Because call quality is very subjective, there are a variety of methods for determining the score, including algorithms that predict MOS scores, which are a popular choice for VoIP networks. Regardless, in a network of reasonable scale, human interaction is perhaps the most efficient, but not necessarily the most practicable, way to assess MOS. The ultimate MOS score is a composite of the participants’ scores, ranging from 0 to 5, with 5 denoting exceptional quality and 0 denoting indecipherable.

Modern testing, on either hand, frequently relies on programs that emphasize on modem reaction times, codec speed, and complicated tests to forecast how the human ear would perceive speech quality. To measure clarity, latency, packet loss, jitter, and establish a likely MOS score, real audio signals are used. Yes, it’s a rough approximation of the human-based MOS score, but it’s more practical, scalable, and measurable.

Following the collection of data, computations to calculate the R-Factor are carried out. These are call-quality-degrading measures that aren’t caused by network errors. Those include the time it takes transmitted data to traverse across the system, the time it takes to decode the signal, and the time it takes to gather voice packets so they may be processed at more periodic intervals.

Three separate metrics are reported in the score:

  • Quality of listening and transmission
  • Quality of Conversation

Testing might concentrate on any and all three metrics at the same time.  A set of people monitors requests for a regulated sample during testing, scoring each measure separately. Measurements are taken after the connection has been tested to identify how to enhance the network’s latency or one-way delay. An algorithm is utilized to grade voice signals in automated testing in an effort to estimate how a human ear would assess the quality and forecast a MOS score.


It’s vital to remember that the MOS is based on a relative scale that takes into account a number of variables that might influence voice quality. As a result, numerous circumstances that would not impact the MOS score on a traditional phone line might affect the MOS score on a VoIP system. These are some of them:

  • Jitter
  • Latency
  • Hardware
  • Bandwidth

One aspect that might have a significant influence on the MOS is the codec version. Compression can impair voice quality dramatically based on the compression ratio. The best voice quality comes from a non-compressed codec, as the signal is less sensitive to audio quality loss due to the absence of compression and decompression.

Despite the fact that compression/decompression systems may be adjusted to save bandwidth, speech quality suffers as a result. The best VoIP codecs for rising MOS would be those that save the most bandwidth while sacrificing the least amount of voice quality. Modern codecs that provide HD-Voice can also tip the scales in the opposite direction, allowing for enhanced quality at the expense of bandwidth economy.


The Mean Opinion Score is a critical tool for VoIP providers and customers to maintain high-quality service, maximum sound quality, and to give technicians the information they need to improve VoIP call voice quality. The score is a subjective or objective measurement of voice clarity, depending on the technique utilized.